README

******************************************

This file contains instructions for replicating "Countering Misperceptions to Reduce Prejudice: An Experiment on Attitudes toward Muslim Americans."
Scott Williamson
Journal of Experimental Political Science
April 2019

******************************************
 
Replication Files:

1. Muslim_Prejudice_Data_SRW: the full dataset used for the primary analysis. This data was collected by Qualtrics on three separate omnibus surveys. The dataset was constructed by downloading the surveys from Qualtrics, limiting each 
spreadsheet to the variables for this project, and then merging them into a single file. 

2. Prejudice_Analysis_JEPS_Final_Williamson: the Stata code to replicate all results in the paper and appendix. 

3. Log_File_Data: the log file for the Stata code to reproduce results in the paper and appendix.

4. Prejudice_Plots_JEPS_Final_Williamson: the R code to replicate the plots in the paper and appendix.

5. Log_File_Plots: the log file for the R code to replicate the plots in the paper and appendix.

6. Williamson_JEPS_Figure1_Data: results from main effects analysis, used as input for Figure 1 in the paper. Constructed in code from "Prejudice_Analysis_JEPS_Final_Williamson."

7. Robust_data_terror: results from terrorism robustness checks, used as input for Figure 2 in the paper. Constructed manually from output in the following section from "Prejudice_Analysis_JEPS_Final_Williamson": "Appendix":
"Terrorism Prime Analysis".

8. Robust_data_sd: results from non-PC license robustness checks, used as input for Figure 2 in the paper. Constructed manually from output in the following section from "Prejudice_Analysis_JEPS_Final_Williamson": "Appendix":
"Non-PC License Analysis."

9: Hetero_coef_data_elderly: results from elderly x treatment interaction, used as input for Figure 3 in the paper. Constructed manually from output in the following section from "Prejudice_Analysis_JEPS_Final_Williamson":
"Main Paper Analysis": "Heterogeneous Results from Figure 3: Elderly Interaction."

10: Hetero_coef_data_white: results from white x treatment interaction, used as input for Figure 3 in the paper. Constructed manually from output in the following section from "Prejudice_Analysis_JEPS_Final_Williamson":
"Main Paper Analysis": "Heterogeneous Results from Figure 3: White Interaction."

11: Hetero_coef_data_no_muslim: results from knows no Muslim x treatment interaction, used as input for Figure 3 in the paper. Constructed manually from output in the following section from "Prejudice_Analysis_JEPS_Final_Williamson":
"Main Paper Analysis": "Heterogeneous Results from Figure 3: Knows No Muslim Interaction."

12: Hetero_coef_data_republican: results from republican x treatment interaction, used as input for Figure 3 in the paper. Constructed manually from output in the following section from "Prejudice_Analysis_JEPS_Final_Williamson":
"Main Paper Analysis": "Heterogeneous Results from Figure 3: Republican Interaction."

13: Eigenvalues: contains the eigenvalues from PCA analysis, used as input for Figure 1 in the appendix. Constructed manually from output in the following section from "Prejudice_Analysis_JEPS_Final_Williamson": "Prepping Dataset: 
"Creating Outcome Variables: PCA." 

******************************************

Codebook for "Muslim_Prejudice_Data_SRW"

Note that language for all survey questions is contained in Section 3 of the Online Appendix

* treatment_srw: exposure to information treatment; 0 = control group, 1 = base treatment group, 2 = treatment group with additional shaming sentence.
* terrorism: exposure to terrorism prime; 0 = no exposure, 1 = exposure
* srw_therm_1_h_1: responses to feeling thermometer outcome without non-PC license sentence in instructions; range = 0 to 100
* srw_therm_2_h_1: responses to feeling thermometer outcome with  non-PC license sentence in instructions, range = 0 to 100
* srw_base_info_h_1: coded 1 if respondent acknowledged surprise at information about Muslim American hobbies in the base treatment group
* srw_base_info_h_2: coded 1 if respondent acknowledged surprise at information about Muslim American attitudes toward political violence in the base treatment group
* srw_base_info_h_3: coded 1 if respondent acknowledged surprise at information about Muslim American identity in the base treatment group
* srw_base_info_h_4: coded 1 if respondent claimed no surprise at Muslim Americans' hobbies, attitudes toward political violence, or identity in the base treatment group
* srw_shame_info_h_1: coded 1 if respondent acknowledged surprise at information about Muslim American hobbies in the shaming treatment group
* srw_shame_info_h_2: coded 1 if respondent acknowledged surprise at information about Muslim American attitudes toward political violence in the shaming treatment group
* srw_shame_info_h_3: coded 1 if respondent acknowledged surprise at information about Muslim American identity in the shaming treatment group
* srw_shame_info_h_4: coded 1 if respondent claimed no surprise at Muslim Americans' hobbies, attitudes toward political violence, or identity in the shaming treatment group
* srw_patriotism: responses to patriotism outcome; 1 = more patriotic, 2 = as patriotic, 3 = less patriotic 
* srw_policy_h_1: responses to surveillance policy outcome; 1 = strongly support, 2 = support, 3 = oppose, 4 = strongly oppose
* srw_policy_h_2: responses to Muslim ban policy outcome; 1 = strongly support, 2 = support, 3 = oppose, 4 = strongly oppose
* srw_policy_h_3: responses to registration policy outcome; 1 = strongly support, 2 = support, 3 = oppose, 4 = strongly oppose
* srw_muslim: coded 1 if respondent claimed to know at least one Muslim in the United States, coded 2 if they did not
* birthyr: respondent's self-reported year of birth
* gender: coded 1 for male, 2 for female
* race: coded 1 for white, 2 for Black or African American, 3 for Hispanic or Latino, 4 for Asian or Asian-American, 5 for Native American, 6 for Middle Eastern, 7 for Mixed Race, 8 for Other
* race_text: write-in responses for racial "other" answer choice
* education: coded 1 for no high school, 2 for high school, 3 for some college, 4 for 2-year degree, 5 for 4-year degree, 6 for postgraduate degree
* ideology: political ideology; coded 1 for very liberal, 2 for liberal, 3 for moderate, 4 for conservative, 5 for very conservative, 6 for not sure
* newsint: news interest; coded 1 for most of the time, 2 for some of the time, 3 for only now and then, 4 for hardly at all, 5 for don't know
* pid_main: political party id; coded 1 for Democrat, 2 for Republican, 3 for neither, 4 for not sure
* pid_dem: among Democrats, coded 1 for strong Democrat, 2 for not very strong Democrat
* pid_rep: among Republicans, coded 1 for strong Republican, 2 for not very strong Republican
* srw_prejudice_h: whether prejudice against minorities is considered an important problem in the United States; 1 = strongly agree, 2 = somewhat agree, 3 = somewhat disagree, 4 = strongly disagree 
* startdate
* enddate
* duration: time to complete survey in seconds
* survey: ID for one of three SSI omnibus surveys; F, G, or H

******************************************

Codebook for "Williamson_JEPS_Figure1_Data"

* information: indicator for exposure to information treatment or not; 0 = control (no information), 1 = treatment (base or shame treatment combined)
* therm: responses to thermometer outcome; 0 to 100
* patriotism: responses to binary patriotism outcome; 0 = less patriotic, 1 = more or equally patriotic
* surveillance: responses to surveillance policy outcomes; 1 to 4
* refugee_ban: responses to Muslim refugee ban policy outcome; 1 to 4
* registration: responses to registration policy outcome; 1 to 4
* pca_attitudes: value for first component from PCA analysis

******************************************

Codebook for spreadsheets that begin with "Robust"

* term: outcome variable
* estimate: coefficient estimate from OLS model 
* std. error: standard error for coefficient
* statistic: t-statistic for coefficient
* p.value: p-value for coefficient
* model: coefficient of interest; treatment effect with terror prime / non-PC license; treatment effect without terror prime / non-PC license; interaction term between terror prime / non-PC license

******************************************

Codebook for spreadsheets that begin with "Hetero"

* term: outcome variable
* estimate: coefficient estimate from OLS models estimating heterogeneous effects 
* std.error: standard error for coefficient
* statistic: t-statistic for coefficient
* p.value: p-value for coefficient
* model: relevant x variables (i.e. relevant demograhpic indicator (elderly, white, knows no Muslim, Republican); treatment indicator; interaction term)

******************************************

Codebook for "Eigenvalues"

* component: components one through five from the principal components analysis
* eigenvalue: the eigenvalue associated with each component